U.S. patent application number 12/885833 was filed with the patent office on 2011-09-22 for spin memory and spin transistor.
This patent application is currently assigned to Kabushiki Kaisha Toshiba. Invention is credited to Hisanori Aikawa, Tomoaki Inokuchi, Mizue Ishikawa, Tatsuya Kishi, Takao Marukame, Masahiko Nakayama, Yoshiaki Saito, Hideyuki Sugiyama, Hiroaki Yoda.
Application Number | 20110228596 12/885833 |
Document ID | / |
Family ID | 44602632 |
Filed Date | 2011-09-22 |
United States Patent
Application |
20110228596 |
Kind Code |
A1 |
Inokuchi; Tomoaki ; et
al. |
September 22, 2011 |
SPIN MEMORY AND SPIN TRANSISTOR
Abstract
Certain embodiments provide a spin memory including a memory
cell including a ferromagnetic stacked film that has a stacked
structure in which a first ferromagnetic layer, a first nonmagnetic
layer, a second ferromagnetic layer, a second nonmagnetic layer,
and a third ferromagnetic layer are stacked in this order or
reverse order, the third ferromagnetic layer and the second
ferromagnetic layer being antiferromagnetically exchange-coupled
via the second nonmagnetic layer. The ferromagnetic stacked film
includes a current path in which a first and second write currents
flow from the first ferromagnetic layer to the third ferromagnetic
layer to write a first and second magnetization states into the
first ferromagnetic layer respectively, and the second write
current is higher than the first write current.
Inventors: |
Inokuchi; Tomoaki;
(Yokohama-Shi, JP) ; Marukame; Takao; (Tokyo,
JP) ; Ishikawa; Mizue; (Yokohama-Shi, JP) ;
Sugiyama; Hideyuki; (Kawasaki-Shi, JP) ; Aikawa;
Hisanori; (Yokohama-Shi, JP) ; Nakayama;
Masahiko; (Yokohama-Shi, JP) ; Kishi; Tatsuya;
(Yokohama-Shi, JP) ; Yoda; Hiroaki;
(Sagamihara-Shi, JP) ; Saito; Yoshiaki;
(Kawasaki-Shi, JP) |
Assignee: |
Kabushiki Kaisha Toshiba
Tokyo
JP
|
Family ID: |
44602632 |
Appl. No.: |
12/885833 |
Filed: |
September 20, 2010 |
Current U.S.
Class: |
365/158 ;
257/295; 257/E29.323 |
Current CPC
Class: |
H01L 43/08 20130101;
G11C 11/1677 20130101; H01L 27/224 20130101; G11C 11/1697 20130101;
G11C 11/1659 20130101; G11C 11/1675 20130101; G11C 11/161 20130101;
H01L 29/66984 20130101 |
Class at
Publication: |
365/158 ;
257/295; 257/E29.323 |
International
Class: |
G11C 11/15 20060101
G11C011/15; H01L 29/82 20060101 H01L029/82 |
Foreign Application Data
Date |
Code |
Application Number |
Mar 19, 2010 |
JP |
2010-65084 |
Claims
1. A spin memory comprising a memory cell including a ferromagnetic
stacked film that has a stacked structure in which a first
ferromagnetic layer, a first nonmagnetic layer, a second
ferromagnetic layer, a second nonmagnetic layer, and a third
ferromagnetic layer are stacked in this order or reverse order, the
third ferromagnetic layer and the second ferromagnetic layer being
antiferromagnetically exchange-coupled via the second nonmagnetic
layer, the ferromagnetic stacked film including a current path in
which a first and second write currents flow from the first
ferromagnetic layer to the third ferromagnetic layer to write a
first and second magnetization states into the first ferromagnetic
layer respectively, and the second write current being higher than
the first write current.
2. The memory according to claim 1, wherein the memory cell further
includes a rectifying device that causes a current to flow in the
ferromagnetic stacked film only in a direction in which the first
and second write currents flow.
3. The memory according to claim 1, wherein the first, second, and
third ferromagnetic layers each have magnetization that is
substantially perpendicular to a film plane.
4. The memory according to claim 1, wherein the first and second
ferromagnetic layers each have a smaller magnetization reversal
magnetic field than a magnetization reversal magnetic field of the
third ferromagnetic layer.
5. The memory according to claim 1, wherein the first write current
makes a magnetization direction of the first ferromagnetic layer
substantially antiparallel to a magnetization direction of the
third ferromagnetic layer, and the second write current reverses a
magnetization direction of the second ferromagnetic layer and makes
the magnetization direction of the first ferromagnetic layer
substantially parallel to the magnetization direction of the third
ferromagnetic layer.
6. The memory according to claim 5, further comprising: a write
circuit that supplies the first and second write currents to the
memory cell.
7. A spin transistor comprising: a source region and a drain region
that are placed at a distance from each other on a semiconductor
layer; a gate electrode that is provided on a portion of the
semiconductor layer, the portion being to serve as a channel region
between the source region and the drain region; a first electrode
that is placed on one of the source region and the drain region,
and has a first stacked structure in which a first ferromagnetic
layer, a first nonmagnetic layer, a second ferromagnetic layer, a
second nonmagnetic layer, and a third ferromagnetic layer are
stacked in this order, the third ferromagnetic layer and the second
ferromagnetic layer being antiferromagnetically exchange-coupled
via the second nonmagnetic layer, the first stacked structure
including a current path in which a first and second write currents
flow from the first ferromagnetic layer to the third ferromagnetic
layer to write a first and second magnetization states into the
first ferromagnetic layer respectively, the second write current
being higher than the first write current; and a second electrode
that is placed on the other one of the source region and the drain
region, and includes a fourth ferromagnetic layer.
8. The transistor according to claim 7, wherein the first, second,
third, and fourth ferromagnetic layers each have an easy axis of
magnetization that is substantially perpendicular to a film
plane.
9. The transistor according to claim 7, further comprising a gate
insulating film that is placed between the channel region and the
gate electrode.
10. The transistor according to claim 7, further comprising: a
first tunnel barrier layer that is placed between the first
ferromagnetic layer and the one of the source region and the drain
region; and a second tunnel barrier layer that is placed between
the fourth ferromagnetic layer and the other one of the source
region and the drain region.
11. The transistor according to claim 7, wherein the second
electrode includes a second stacked structure that includes a third
nonmagnetic layer placed on the fourth ferromagnetic layer, a fifth
ferromagnetic layer placed on the third nonmagnetic layer, a fourth
nonmagnetic layer placed on the fifth ferromagnetic layer, and a
sixth ferromagnetic layer placed on the fourth nonmagnetic layer,
the fifth ferromagnetic layer and the sixth ferromagnetic layer
being antiferromagnetically exchange-coupled via the fourth
nonmagnetic layer, and a film plane of the fourth ferromagnetic
layer has a larger area than a film plane of the first
ferromagnetic layer.
12. A spin transistor comprising: a first electrode that is placed
on a semiconductor substrate, and includes a first ferromagnetic
layer having an easy axis of magnetization that is substantially
perpendicular to a film plane; a semiconductor layer that is placed
on the first ferromagnetic layer, and is to serve as a channel; a
second electrode that has a stacked structure that is placed on the
semiconductor layer, and includes a second ferromagnetic layer
having an easy axis of magnetization substantially perpendicular to
a film plane, a first nonmagnetic layer placed on the second
ferromagnetic layer, a third ferromagnetic layer that is placed on
the first nonmagnetic layer and has an easy axis of magnetization
substantially perpendicular to a film plane, a second nonmagnetic
layer placed on the third ferromagnetic layer, and a fourth
ferromagnetic layer that is placed on the second nonmagnetic layer
and has an easy axis of magnetization substantially perpendicular
to a film plane, the fourth ferromagnetic layer being
antiferromagnetically exchange-coupled to the third ferromagnetic
layer via the second nonmagnetic layer, the stacked structure
including a current path in which a first and second write currents
flow from the second ferromagnetic layer to the fourth
ferromagnetic layer to write a first and second magnetization
states into the second ferromagnetic layer respectively, the second
write current being higher than the first write current; a gate
insulating film that is placed on a side face of the semiconductor
layer; and a gate electrode that is placed on the opposite side of
the gate insulating film from the semiconductor layer.
13. The transistor according to claim 12, further comprising: a
first tunnel barrier layer that is placed between the first
ferromagnetic layer and the semiconductor layer; and a second
tunnel barrier layer that is placed between the semiconductor layer
and the second ferromagnetic layer.
Description
CROSS-REFERENCE TO RELATED APPLICATION
[0001] This application is based upon and claims the benefit of
priority from prior Japanese Patent Application No. 2010-65084
filed on Mar. 19, 2010 in Japan, the entire contents of which are
incorporated herein by reference.
FIELD
[0002] Embodiments described herein relate to a spin memory and a
spin transistor.
BACKGROUND OF THE INVENTION
[0003] In recent years, magnetoresistive devices that utilize
tunnel magnetoresistive effects are actively and experimentally
used in magnetic random access memories.
[0004] Where a magnetoresistive device is used as a memory device
in a magnetic random access memory, the magnetoresistive device
normally has a ferromagnetic single tunnel junction or a
ferromagnetic double tunnel junction. In a ferromagnetic single
tunnel junction, one of the two ferromagnetic layers sandwiching a
tunnel barrier layer is a fixed magnetization layer (a
magnetization reference layer) that has a fixed magnetization
direction, and the other one of the two ferromagnetic layers is a
magnetic recording layer that has a variable magnetization
direction. A ferromagnetic double tunnel junction has a stacked
structure formed with a first magnetization reference layer, a
first tunnel barrier layer, a magnetic recording layer, a second
tunnel barrier layer, and a second magnetization reference layer.
Hereinafter, a ferromagnetic single tunnel junction or a
ferromagnetic double tunnel junction will also be referred to
simply as a MTJ. A memory device having such a MTJ is capable of
storing data in a nonvolatile manner, and characteristically
requires a short writing time or a short reading time of 10 ns or
less. Also, such a memory device is characteristically capable of
performing rewriting 10.sup.15 times or more.
[0005] As a method of performing writing on such a magnetoresistive
device, a spin-injection magnetization reversing technique has been
suggested (see U.S. Pat. No. 6,256,223, for example). By the
spin-injection magnetization reversing technique, the magnetization
of a magnetic recording layer is reversed by injecting
spin-polarized electrons (spin injection current) into a magnetic
recording layer of a memory device. The value of the spin injection
current required for writing becomes smaller as the magnetic
recording layer becomes smaller. Accordingly, there is a high
expectation for the spin-injection magnetization reversing
technique as a method of performing writing on a magnetic random
access memory.
[0006] However, to perform magnetic writing by the spin-injection
magnetization reversing technique, it is necessary to cause a
bidirectional current to flow in a magnetoresistive device.
Therefore, diode-type architecture (so-called cross-point
architecture) that connects magnetoresistive devices and diodes in
series between bit lines and word lines cannot be used. In view of
this, architecture in which each one memory cell is formed with at
least one transistor and one magnetoresistive device needs to be
employed, and, as a result, the cell size cannot be made equal to
or smaller than the cell size of a dynamic random access memory
(DRAM).
[0007] As a memory device of a spin-injection magnetization
reversal type, there is a known spin transistor (hereinafter also
referred to as a spin MOS transistor) having a MTJ in at least
either a source electrode or a drain electrode (see Japanese
Laid-Open Patent Publication No. 2008-66596, for example). Since a
bidirectional current is required at the time of writing in this
spin transistor, a switching transistor for switching the direction
of the current needs to be connected to each of the source and
drain electrodes of the spin transistor. Therefore, the problem of
an increased total circuit area is caused as in a magnetoresistive
device that performs magnetic writing with spin-injection
magnetization reversals.
[0008] As described above, a bidirectional current is required for
writing in a spin memory or a spin transistor of a conventional
spin-injection magnetization reversal type, and the total circuit
area cannot be made smaller.
BRIEF DESCRIPTION OF THE DRAWINGS
[0009] FIG. 1 is a schematic view of a spin memory according to a
first embodiment of the present invention;
[0010] FIG. 2 is a cross-sectional view for explaining a method of
writing a low-resistance state into the spin memory according to
the first embodiment;
[0011] FIG. 3 is a cross-sectional view for explaining a method of
writing a high-resistance state into the spin memory according to
the first embodiment;
[0012] FIG. 4 is a cross-sectional view for explaining a method of
writing a high-resistance state into the spin memory according to
the first embodiment;
[0013] FIG. 5 is a circuit diagram showing a specific example
circuit structure of the spin memory according to the first
embodiment;
[0014] FIG. 6 is a circuit diagram showing a specific example write
circuit and a specific example read circuit of the spin memory
according to the first embodiment;
[0015] FIG. 7 is a cross-sectional view of a spin transistor
according to a second embodiment;
[0016] FIG. 8 is a cross-sectional view of a spin transistor
according to a first modification of the second embodiment;
[0017] FIG. 9 is a cross-sectional view of a spin transistor
according to a second modification of the second embodiment;
[0018] FIG. 10 is a cross-sectional view of a spin transistor
according to a third modification of the second embodiment;
[0019] FIG. 11 is a cross-sectional view of a spin transistor
according to a fourth modification of the second embodiment;
[0020] FIG. 12 is a cross-sectional view of a spin transistor
according to a fifth modification of the second embodiment;
[0021] FIG. 13 is a cross-sectional view of a spin transistor
according to a third embodiment; and
[0022] FIG. 14 is a cross-sectional view of a spin transistor
according to a modification of the third embodiment.
DETAILED DESCRIPTION
[0023] Certain embodiments provide a spin memory including a memory
cell including a ferromagnetic stacked film that has a stacked
structure in which a first ferromagnetic layer, a first nonmagnetic
layer, a second ferromagnetic layer, a second nonmagnetic layer,
and a third ferromagnetic layer are stacked in this order or
reverse order, the third ferromagnetic layer and the second
ferromagnetic layer being antiferromagnetically exchange-coupled
via the second nonmagnetic layer. The ferromagnetic stacked film
includes a current path in which a first and second write currents
flow from the first ferromagnetic layer to the third ferromagnetic
layer to write a first and second magnetization states into the
first ferromagnetic layer respectively, and the second write
current is higher than the first write current.
[0024] Hereafter, a spin memory and a spin transistor according to
embodiments invention will be described more specifically with
reference to the drawings.
First Embodiment
[0025] FIG. 1 shows a spin memory according to a first embodiment
of the present invention. The spin memory of the first embodiment
has memory cells 10 (two memory cells 10 in FIG. 1) arranged in a
matrix fashion. Each of the memory cells 10 includes a
ferromagnetic stacked film 20 and a diode 30. The ferromagnetic
stacked film 20 has a stacked structure in which a ferromagnetic
layer 21, a nonmagnetic layer 23, a ferromagnetic layer 25, a
nonmagnetic layer 27, and a ferromagnetic layer 29 are stacked in
this order. The ferromagnetic layer 21, the ferromagnetic layer 25,
and the ferromagnetic layer 29 each have an easy axis of
magnetization that is substantially perpendicular to the film
plane. Specifically, when there is not an external magnetic field,
the magnetization direction of each of the ferromagnetic layers is
substantially perpendicular to the film plane. Here, the film plane
means the upper face of each ferromagnetic layer. The magnetization
reversal magnetic fields of the ferromagnetic layers 21 and 25 are
smaller than that of the ferromagnetic layer 29. The ferromagnetic
layer 29 and the ferromagnetic layer 25 are antiferromagnetically
magnetic-exchange-coupled via the nonmagnetic layer 27, and the
magnetization direction of the ferromagnetic layer 29 and the
magnetization direction of the ferromagnetic layer 25 are the
opposite of (antiparallel to) each other. Accordingly, where the
magnetization direction of the ferromagnetic layer 29 is an upward
direction, the magnetization direction of the ferromagnetic layer
25 is a downward direction, as shown in FIG. 1.
[0026] The anode of the diode 30 is electrically connected to a
word line WL, and the cathode is electrically connected to the
ferromagnetic layer 21. The ferromagnetic layer 29 is connected to
a bit line BL. In other words, each memory cell 10 is architecture
of a cross-point type that has one end connected to a word line WL
and has the other end connected to a bit line BL. To reduce contact
resistance, a low-resistance barrier metal, such as a metal layer
made of Ti, TiN, W, or the like, may be inserted between the anode
of the diode 30 and the word line WL, and between the cathode and
the ferromagnetic layer 21.
[0027] Although the anode of each diode 30 is connected to the word
line WL and the cathode of each diode 30 is connected to the
ferromagnetic layer 21 in FIG. 1, each diode 30 may be connected to
the side of the bit line BL. Specifically, the cathode of each
diode 30 may be connected to the bit line BL, and the anode may be
connected to the ferromagnetic layer 29.
[0028] Referring now to FIGS. 2 through 4, a method of performing
writing in a memory cell of the spin memory according to the first
embodiment is described.
[0029] First, as shown in FIG. 2, when electrons having a current
of a magnitude I1 (>0) are caused to flow from the ferromagnetic
layer 29 to the ferromagnetic layer 21, the relative magnetization
directions of the ferromagnetic layer 21 and the ferromagnetic
layer 25 become substantially parallel to each other, based on the
principles of spin-injection magnetization reversals (FIG. 2). In
FIG. 2, the arrows indicate the directions of the flows of
electrons, and the current I1 flows in the opposite direction from
the arrows. At this point, the magnitude of the current I1 is such
a magnitude as not to overcome the ferromagnetic magnetic-exchange
coupling between the ferromagnetic layer 25 and the ferromagnetic
layer 29. If a current I2 (>0) that is sufficiently smaller than
(or does not cause a spin-injection magnetization reversal with)
the magnitude of the current I1 is caused to flow in this
situation, and the memory cell resistance is read, the relative
magnetization directions of the ferromagnetic layer 25 and the
ferromagnetic layer 21 become substantially parallel to each other.
Accordingly, the ferromagnetic stacked film 20 or the memory cell
10 is put into a low-resistance state.
[0030] Next, a case where writing is performed with a current of a
greater magnitude than I1 is described.
[0031] Since the ferromagnetic layer 29 and the ferromagnetic layer
25 are in contact with each other via the nonmagnetic layer 27 in
the present situation, a spin-injection magnetization reversal
between the ferromagnetic layer 29 and the ferromagnetic layer 25
can be caused by selecting appropriate materials for the
ferromagnetic layers 29 and 25 and the nonmagnetic layer 27, and
adjusting the thickness of the nonmagnetic layer 27 to an
appropriate thickness. When a larger current than the current I1
flows between the ferromagnetic layer 29 and the ferromagnetic
layer 21, or a current I3 of such a magnitude (>0) as to cause a
spin-injection magnetization reversal in the ferromagnetic layer 25
flows, a spin-injection magnetization reversal is induced between
the ferromagnetic layer 29 and the ferromagnetic layer 25, and the
relative magnetization directions of the ferromagnetic layer 29 and
the ferromagnetic layer 25 become substantially parallel to (the
same as) each other. In FIG. 3, the arrows indicate the directions
of the flows of electrons, and the current I3 flows in the opposite
direction from the arrows. Since a spin-injection magnetization
reversal is induced between the ferromagnetic layer 25 and the
ferromagnetic layer 21 via the nonmagnetic layer 23, the relative
magnetization directions of the ferromagnetic layer 25 and the
ferromagnetic layer 21 become substantially antiparallel to each
other (FIG. 3). In the above description, the current I3 is defined
as a current of such a magnitude as to cause a spin-injection
magnetization reversal in the ferromagnetic layer 25, but may be
defined as a current of such a magnitude as to convert the
antiferromagnetic magnetic-exchange coupling between the
ferromagnetic layer 29 and the ferromagnetic layer 25 to
ferromagnetic magnetic-exchange coupling.
[0032] Next, a case where the current is shut off in the situation
illustrated in FIG. 3 is described. Since there is substantially
antiparallel magnetic-exchange coupling between the ferromagnetic
layer 29 and the ferromagnetic layer 25 via the nonmagnetic layer
27, the relative magnetization directions of the ferromagnetic
layer 29 and the ferromagnetic layer 25 become substantially
antiparallel to each other when the current stops flowing. In other
words, when the current is shut off, the magnetization direction of
the ferromagnetic layer 25 is reversed or becomes an opposite
direction of the previous direction. Therefore, as shown in FIG. 4,
the relative magnetization directions of the ferromagnetic layer 25
and the ferromagnetic layer 29 become substantially antiparallel to
each other. At this point, there is not a current flowing in the
ferromagnetic stacked film 20, and therefore, the magnetization
direction of the ferromagnetic layer 21 does not change. If the
current I2 that is sufficiently smaller than (or does not cause a
spin-injection magnetization reversal with) the magnitude of the
current I1 is caused to flow in this situation, and the memory cell
resistance is read, the relative magnetization directions of the
ferromagnetic layer 25 and the ferromagnetic layer 21 become
substantially antiparallel to each other. Accordingly, the
ferromagnetic stacked film 20 or the memory cell 10 is put into a
high-resistance state.
[0033] As described above, the magnetized state of the
ferromagnetic layer 21 in the memory cell 10 is rewritten with the
use of only unidirectional currents. In other words, the memory
cell 10 can be rewritten to be in a low-resistance state or a
high-resistance state.
[0034] By appropriately selecting the materials of the
ferromagnetic layer 21 and the ferromagnetic layer 25, the relative
magnetization directions of the ferromagnetic layer 21 and the
ferromagnetic layer 25 after the write current I1 or I3 is supplied
can be set in a high-resistance state or a low-resistance state.
For example, different kinds of magnetic metal materials are used
for the ferromagnetic layer 21 and the ferromagnetic layer 25. In
the combination that may be used here, the minority state density
is high in one of the magnetic metal materials in terms of Fermi
level, and the majority state density is high in the other.
[0035] To determine whether desired writing has been correctly
performed, a verify write method may be used. Specifically, after
writing is performed with the use of the write current I1 or I3,
the magnetization direction of the memory cell on which the writing
has been performed is read with the use of a lower current than the
write currents I1 and I3. A check is then made to determine whether
the desired writing has been performed. If the desired writing has
not been performed, the desired writing may be performed again with
the use of the above described write current. If the desired
writing has not been performed, however, writing may be performed
with a greater write pulse width, or writing may be performed
several times, so as to increase the probability that the desired
writing is properly performed. Alternatively, depending on the
result of reading, writing may be again performed with a higher or
lower write current. For example, if writing cannot be performed
with the current I3, writing is again performed with a higher
current than the current I3. If writing cannot be performed with
the current I1, writing is again performed with a lower current
than the current I1.
[0036] As described above, according to this embodiment, writing
can be performed with unidirectional currents, and an increase of
circuit area can be prevented by virtue of the cross-point type
architecture.
[0037] In this embodiment, each memory cell 10 includes a diode 30.
This diode 30 is used to allow only unidirectional currents or is
used for rectification. Therefore, another element (a rectifying
device) having a rectifying function, such as a Coulomb blockade
device, may be used, instead of a diode.
[0038] The rectifying device such as the diode 30 may be omitted
from each memory cell 10. In this case, there is a possibility that
the current might enter from surrounding memory cells via bit lines
and word lines. Therefore, it is necessary to increase the
resistance change rate or the difference in resistance value
between a low-resistance state and a high-resistance state.
[0039] FIG. 5 shows a specific example circuit structure of the
spin memory according to the first embodiment. This spin memory
includes memory cells 10 arranged in a matrix fashion, word lines
WL, bit lines BL intersecting with those word lines WL, a
row-decoder/write-circuit/read-circuit unit 100, and a
column-decoder/read-circuit unit 150. Each of the memory cells 10
is placed near an intersecting area between the word lines WL and
the bit lines BL, and includes the diode 30 and the ferromagnetic
stacked film 20 explained with reference to FIG. 1.
[0040] FIG. 6 shows a specific example of the
row-decoder/write-circuit/read-circuit unit 100 and the
column-decoder/read-circuit unit 150. The
row-decoder/write-circuit/read-circuit unit 100 includes three
current sources 101.sub.1, 101.sub.2, and 101.sub.3, three switches
103.sub.1, 103.sub.2, and 103.sub.3 formed with p-type MOS
transistors, for example, and a select transistor 110 that is
provided in a corresponding word line WL and is formed with an
n-type MOS transistor, for example. The intensities of the currents
generated from the current sources 101.sub.1, 101.sub.2, and
101.sub.3 are I1, I2, and I3, respectively. Here, the relationships
among the current intensities are expressed as: I2<I1<I3. The
current source 101.sub.1 and the switch 103.sub.1 are connected in
series, and form a first series circuit. The current source
101.sub.2 and the switch 103.sub.2 are connected in series, and
form a second series circuit. The current source 101.sub.3 and the
switch 103.sub.3 are connected in series, and form a third series
circuit. The first series circuit and the third series circuit form
a write circuit, and the second series circuit forms a read
circuit. Each of the first through third series circuits is
connected to one end of the select transistor 110. The other end of
the select transistor 110 is connected to the corresponding word
line WL.
[0041] The column-decoder/read-circuit 150 includes a select
transistor 152 that is provided in a corresponding bit line BL and
is formed with an n-type MOS transistor, and a resistor 154. One
end of the select transistor 152 is connected to the corresponding
bit line BL, and the other end is grounded via the resistor
154.
[0042] The operations of the spin memory of the first embodiment
having the above-described structure are now described.
[0043] First, writing to be performed to cause the ferromagnetic
layer 21 of a memory cell 10 to have the magnetization direction
shown in FIG. 2 with the use of a current having the intensity I1
is described. In this case, the
row-decoder/write-circuit/read-circuit unit 100 sends a control
signal to the gate of the select transistor 110, to turn on the
select transistor 110. The column-decoder/read-circuit unit 150
sends a control signal to the gate of the select transistor 152, to
turn on the select transistor 152. In this manner, the memory cell
10 on which the writing is to be performed is selected. The
row-decoder/write-circuit/read-circuit unit 100 then sends a
control signal to the gate of the switch 103.sub.1, to turn on the
switch 103.sub.1. The write current I1 then flows from the current
source 101.sub.1 into the selected memory cell 10 through the
switch 103.sub.1, and the writing is performed to cause the
selected memory cell 10 to have the magnetization directions shown
in FIG. 2. This writing is ended when the
row-decoder/write-circuit/read-circuit unit 100 sends a control
signal to the gate of the switch 103.sub.1 and turns off the switch
103.sub.1.
[0044] Next, writing to be performed to cause the ferromagnetic
layer 21 of a memory cell 10 to have the magnetization direction
shown in FIG. 4 with the use of a current having the intensity I3
is described. The memory cell 10 on which the writing is to be
performed is selected in the same manner as above. The
row-decoder/write-circuit/read-circuit unit 100 then sends a
control signal to the gate of the switch 103.sub.3, to turn on the
switch 103.sub.3. The write current I3 then flows from the current
source 101.sub.3 into the selected memory cell 10 through the
switch 103.sub.3, and the writing is performed to cause the
selected memory cell 10 to have the magnetization directions shown
in FIG. 4. This writing is ended when the
row-decoder/write-circuit/read-circuit unit 100 sends a control
signal to the gate of the switch 103.sub.3 and turns off the switch
103.sub.3.
[0045] Next, a reading operation is described. The memory cell 10
on which reading is to be performed is selected in the same manner
as above. The row-decoder/write-circuit/read-circuit unit 100 then
sends a control signal to the gate of the switch 103.sub.2, to turn
on the switch 103.sub.2. The write current I2 then flows from the
current source 101.sub.2 into the selected memory cell 10 through
the switch 103.sub.2. At this point, a voltage corresponding to the
magnetization direction of the ferromagnetic layer 21 of the
selected memory cell 10 appears at a connecting node 153 between
the select transistor 152 and the resistor 154. By reading the
voltage of the connecting node 153, a check can be made to
determine whether the selected memory cell 10 is in a
high-resistance state or a low-resistance state. The device for
reading the voltage of the connecting node 153 and the device for
determining whether the selected memory cell 10 is in a
high-resistance state or a low-resistance state based on the
voltage are not shown in FIG. 6, but are included in the
column-decoder/read-circuit unit 150.
[0046] As described above, according to the first embodiment,
writing can be performed with unidirectional currents, and an
increase of the circuit area can be prevented.
Second Embodiment
[0047] FIG. 7 shows a spin transistor according to a second
embodiment of the present invention. The spin transistor of the
second embodiment is a spin transistor that is formed on a
semiconductor layer 41. Here, a "semiconductor layer" means a
semiconductor substrate, a well region formed in a semiconductor
substrate, a SOI (Silicon On Insulator) layer in a SOI substrate,
or the like. An impurity region 43a and an impurity region 43b that
are doped with impurities at high density are provided at a
distance from each other on the surface of the semiconductor layer
41. In this embodiment, the semiconductor layer 41 is an n-type
semiconductor layer, the impurity region 43a and the impurity
region 43b are a p-type impurity region 43a and a p-type impurity
region 43b, and the spin transistor is a p-type spin transistor. In
this case, the p-type impurity region 43a and the p-type impurity
region 43b are to be a source region 43a and a drain region 43b. A
channel region 42 is formed in a portion of the semiconductor layer
41 located between the source region 43a and the drain region 43b.
A gate insulating film 45 is provided on the channel region 42, and
a gate electrode 47 is provided on the gate insulating film 45. As
the source region 43a and the drain region 43b, a
semiconductor-metal compound (such as silicide) may be used,
instead of a material doped with high-density impurities.
[0048] A source electrode 50 is provided on the source region 43a,
and a drain electrode 60 is provided on the drain region 43b. The
source electrode 50 has a first ferromagnetic stacked film in which
a nonmagnetic layer 51 and a ferromagnetic layer 52 are stacked in
this order. The drain electrode 60 has a second ferromagnetic
stacked film in which a nonmagnetic layer 61, a ferromagnetic layer
62, a nonmagnetic layer 63, a ferromagnetic layer 64, a nonmagnetic
layer 65, and a ferromagnetic layer 66 are stacked in this order.
Accordingly, the ferromagnetic layer 52 and the ferromagnetic layer
62 are electrically connected to each other via the nonmagnetic
layer 51, the source region 43a, the channel region 42, the drain
region 43b, and the nonmagnetic layer 61.
[0049] The nonmagnetic layer 51 and the nonmagnetic layer 61 serve
to prevent a reaction that might occur between the surface of the
semiconductor layer 41 and the ferromagnetic layer 52, and a
reaction that might occur between the surface of the semiconductor
layer 41 and the ferromagnetic layer 62, respectively. Accordingly,
the junction between the ferromagnetic layer 52 and the surface of
the semiconductor layer 41, and the junction between the
ferromagnetic layer 62 and the surface of the semiconductor layer
41 can be properly formed, and the transistor manufacture yield can
be improved. Further, by virtue of the nonmagnetic layer 51 and the
nonmagnetic layer 61, carriers can be injected into the channel
region 42 at a high spin polarization rate. As a result, the
conductance can be caused to vary more widely with the changes of
the magnetization directions of the ferromagnetic layer 52 and the
ferromagnetic layer 62. Depending on the performance required for
the transistor, it is possible to form a nonmagnetic layer only
between the surface of the semiconductor layer 41 and the
ferromagnetic layer 62, or only between the surface of the
semiconductor layer 41 and the ferromagnetic layer 52.
[0050] The ferromagnetic layer 52, the ferromagnetic layer 62, the
ferromagnetic layer 64, and the ferromagnetic layer 66 each have an
easy axis of magnetization that is substantially perpendicular to
the film plane. More specifically, when there is not an external
magnetic field, the magnetization direction of each of the
ferromagnetic layers is substantially perpendicular to the film
plane. The magnetization reversal magnetic fields of the
ferromagnetic layers 62 and 64 are smaller than those of the
ferromagnetic layers 52 and 66. The ferromagnetic layer 66 and the
ferromagnetic layer 64 form antiferromagnetic magnetic-exchange
coupling via the nonmagnetic layer 65, and the magnetization
direction of the ferromagnetic layer 66 and the magnetization
direction of the ferromagnetic layer 64 are the opposite from
(antiparallel to) each other.
[0051] In the spin transistor of the second embodiment, the current
obtained when a voltage is applied between the ferromagnetic layer
52 and the ferromagnetic layer 66 depends on the voltage applied to
the gate electrode 47 and the relative magnetization directions of
the ferromagnetic layer 52 and the ferromagnetic layer 62.
Specifically, if the relative magnetization directions of the
ferromagnetic layer 52 and the ferromagnetic layer 62 are
substantially parallel to each other when a gate voltage is
applied, a large current flows. If the relative magnetization
directions of the ferromagnetic layer 52 and the ferromagnetic
layer 62 are substantially antiparallel to each other when a gate
voltage is applied, the amount of current to flow is small.
[0052] In the spin transistor of the second embodiment, the
magnetization direction of the ferromagnetic layer 62 can be
changed by causing currents with different intensities to flow from
the first ferromagnetic stacked film 50 to the second ferromagnetic
stacked film 60 through the channel region 42 (in the opposite
direction from the flow of electrons), as described in the first
embodiment. In other words, the magnetization direction of the
ferromagnetic layer 62 can be changed only with unidirectional
currents, as described in the first embodiment. In the p-type spin
transistor of this embodiment, currents are caused to flow from the
first ferromagnetic stacked film 50 serving as the source electrode
to the second ferromagnetic stacked film 60 serving as the drain
electrode through the channel region 42. In an n-type spin
transistor, however, if the semiconductor layer 41 is a p-type
semiconductor layer while the impurity regions 43a and 43b are
n-type impurity regions, the impurity region 43a and the first
ferromagnetic stacked film 50 turn into the drain region and the
drain electrode, respectively, and the impurity region 43b and the
second ferromagnetic stacked film 60 turn into the source region
and the source electrode, respectively.
[0053] Unlike the spin transistor of this embodiment, the spin
transistor disclosed in JP-A 2008-66596 (KOKAI), for example,
requires bidirectional currents at the time of writing, and
therefore, has problems such as large-size peripheral circuits and
limited locations of use.
[0054] In the spin transistor of this embodiment, on the other
hand, writing with unidirectional currents can be performed.
Accordingly, the peripheral circuits can be simplified, and high
compatibility with regular logic circuits can be secured.
[0055] Instead of the first stacked film 50 provided as the source
electrode 50, a stacked film 60A having the same structure as the
second stacked film 60 provided as the drain electrode 60 may be
provided as in a first modification illustrated in FIG. 8. More
specifically, the stacked film 60A has a stacked structure in which
a nonmagnetic layer 61A, a ferromagnetic layer 62A, a nonmagnetic
layer 63A, a ferromagnetic layer 64A, a nonmagnetic layer 65A, and
a ferromagnetic layer 66A are stacked in this order. In this case,
the stacked film 60A as the source electrode is preferably made
larger than the second stacked film 60 as the drain electrode, so
that the magnetization of the ferromagnetic layer 62A is not
reversed when a current is caused to flow from the stacked film 60A
to the stacked film 60 through the channel region 42 (in the
opposite direction from the flow of electrons).
[0056] This embodiment and the first modification illustrated in
FIG. 7 and FIG. 8 concern MOSFET structures each having the gate
insulating film 45. However, it is possible to use a MESFET
structure that does not have the gate insulating film 45 as in
second and third modifications illustrated in FIG. 9 and FIG.
10.
[0057] In this embodiment and the modifications illustrated in
FIGS. 7 through 10, the source electrode 50 has the nonmagnetic
layer provided between the semiconductor layer 41 and the
ferromagnetic layer 52, and the drain electrode 60 has the
nonmagnetic layer 61 provided between the semiconductor layer 41
and the ferromagnetic layer 62. However, a source electrode 50A
from which the nonmagnetic layer 51 is eliminated, and a drain
electrode 60A from which the magnetic layer 61 is eliminated may be
used as in fourth and fifth modifications illustrated in FIG. 11
and FIG. 12.
[0058] As described above, according to the second embodiment and
the modifications, writing can be performed with unidirectional
currents, and there is no need to use a transistor for switching
the directions of currents. Accordingly, the circuit area can be
prevented from increasing.
[0059] In the second embodiment, a verify write method may be used
to determine whether desired writing has been correctly performed,
as in the first embodiment. If the desired writing has not been
performed, writing may be performed with a greater write pulse
width, or writing may be performed several times, so as to increase
the probability that the desired writing is properly performed.
Alternatively, depending on the result of reading, writing may be
again performed with a higher or lower write current. For example,
if writing cannot be performed with the current I3, writing is
again performed with a higher current than the current I3. If
writing cannot be performed with the current I1, writing is again
performed with a lower current than the current I1.
Third Embodiment
[0060] Next, a spin transistor according to a third embodiment of
the present invention is described. The spin transistor of the
second embodiment illustrated in FIG. 7 is of a transverse type,
and has a structure in which the source electrode and the source
region having ferromagnetic layers are placed on a side of the
channel region, and the drain region and the drain electrode are
placed on the other side of the channel region. The spin transistor
of the third embodiment is of a longitudinal type, and has a
structure in which a source electrode, a channel region, and a
drain electrode are stacked in this order on a semiconductor
layer.
[0061] FIG. 13 shows the spin transistor of the third embodiment.
The spin transistor of this third embodiment is placed on a SOI
substrate 70 that includes a supporting substrate 71 made of Si, a
buried oxide film 72, and a SOI layer 73. The spin transistor of
the third embodiment may be placed on a bulk substrate made of Si,
instead of a SOI substrate. A conductive base layer 75 is provided
on the SOI layer 73. A stacked structure in which a source
electrode 80, a semiconductor crystalline layer 84 to be a channel
region, and a drain electrode 90 are stacked in this order is
provided on the base layer 75. Alternatively, a drain electrode, a
channel region, and a source electrode may be stacked in this order
on the base layer. A gate electrode 86 is formed to surround the
semiconductor crystalline layer 84 serving as the channel region of
the stacked structure, and has a gate insulating film 88a
interposed between the semiconductor crystalline layer 84 and the
gate electrode 86. In this embodiment, the gate electrode 86 is
formed to surround the semiconductor crystalline layer 84. However,
the gate insulating film 88a may be interposed between the gate
electrode 86 and part of the side faces of the semiconductor
crystalline layer 84. An insulating film 88 also exists between the
gate electrode 86 and the base layer 75.
[0062] In this embodiment, the spin transistor is an n-type spin
transistor, and the semiconductor crystalline layer 84 is a p-type
semiconductor crystalline layer. In a p-type spin transistor in
which the semiconductor crystalline layer 84 is an n-type
semiconductor crystalline layer, the electrode 80 and the electrode
90 shown in FIG. 13 are the drain electrode and the source
electrode, respectively.
[0063] The source electrode 80 includes a ferromagnetic layer 81
provided on the base layer 75, and a nonmagnetic layer 82 provided
on the ferromagnetic layer 81. The drain electrode 90 has a stacked
structure in which a nonmagnetic layer 91, a ferromagnetic layer
92, a nonmagnetic layer 93, a ferromagnetic layer 94, a nonmagnetic
layer 95, and a ferromagnetic layer 96 are stacked in this order on
the semiconductor crystalline layer 84. The ferromagnetic layer 81,
the ferromagnetic layer 92, the ferromagnetic layer 94, and the
ferromagnetic layer 96 each have an easy axis of magnetization that
is substantially perpendicular to the film plane. Specifically,
when there is not an external magnetic field, the magnetization
direction of each of the ferromagnetic layers is substantially
perpendicular to the film plane. The magnetization reversal
magnetic fields of the ferromagnetic layers 92 and 94 are smaller
than those of the ferromagnetic layers 81 and 96. The ferromagnetic
layer 96 and the ferromagnetic layer 94 form antiferromagnetic
magnetic-exchange coupling via the nonmagnetic layer 95, and the
magnetization direction of the ferromagnetic layer 96 and the
magnetization direction of the ferromagnetic layer 94 are the
opposite from (antiparallel to) each other.
[0064] In the spin transistor of the third embodiment, the current
obtained when a voltage is applied between the ferromagnetic layer
81 and the ferromagnetic layer 96, or between the base layer 75 and
the ferromagnetic layer 96, depends on the voltage applied to the
gate electrode 86 and the relative magnetization directions of the
ferromagnetic layer 81 and the ferromagnetic layer 92.
Specifically, if the relative magnetization directions of the
ferromagnetic layer 81 and the ferromagnetic layer 92 are
substantially parallel to each other when a gate voltage is
applied, a large current flows. If the relative magnetization
directions of the ferromagnetic layer 81 and the ferromagnetic
layer 92 are substantially antiparallel to each other when a gate
voltage is applied, the amount of current to flow is small.
[0065] In the spin transistor of the third embodiment, the
magnetization direction of the ferromagnetic layer 92 can be
changed only with unidirectional currents, as described in the
second embodiment.
[0066] In the third embodiment, a source region and a drain region
that are doped with high-density impurities of the opposite
conductivity type from the semiconductor crystalline layer 84
should preferably be formed on the semiconductor crystalline layer
side of the interface between the nonmagnetic layer 82 and the
semiconductor crystalline layer 84, and on the semiconductor
crystalline layer side of the interface between the nonmagnetic
layer 91 and the semiconductor crystalline layer 84.
[0067] As the source electrode 80, a ferromagnetic stacked film
having the same structure as the drain electrode 90 may be used. In
this case, the layers in the ferromagnetic stacked film as the
source electrode are stacked in the reverse order of the layers in
the ferromagnetic stacked film as the drain electrode. Also, each
layer in the ferromagnetic stacked film as the source electrode
should preferably have a large film thickness or area to maintain a
large volume so that the magnetization directions are not reversed
when a current is caused to flow.
[0068] In the third embodiment, the SOI layer 73 of the lowermost
layer may be a Si crystalline layer or a SiGe crystalline layer.
The material of the semiconductor crystalline layer 84 may be Ge,
Si, SiGe, GaAs, InGaAs, or the like. Here, it is essential to take
into consideration the influence the crystal lattice of the SOI
layer 73 as the lowermost layer has on the lattice spacing in the
ferromagnetic layer 81 and the semiconductor crystalline layer 84.
Particularly, the lattice and strain of the semiconductor
crystalline layer 84 have a large influence on the mobility of the
carriers in the semiconductor crystalline layer 84. The lattice
spacing in the ferromagnetic layer 81 should preferably be matched
to the lattice of the base layer 75. However, the crystal lattice
of the semiconductor crystalline layer 84 is made slightly
different from the crystal lattice of the base layer 75, and strain
is actively caused. In this manner, the channel mobility can be
improved. For example, SiGe crystals of 80% in Ge composition ratio
are placed as the base layer 75, and the ferromagnetic layer 81
lattice-matched to those crystals is stacked. A Ge crystalline
layer is then stacked as the semiconductor crystalline layer 84
that is to be the channel region. The Ge crystal lattice of the
semiconductor crystalline layer 84 then becomes slightly larger
than the lattice of the SiGe crystal layer as the base layer 75.
Accordingly, the Ge crystal lattice of the semiconductor
crystalline layer 84 is subjected to compression strain and shrinks
in the horizontal direction to be matched to the base layer 75, and
is subjected to tension stress and expands in the vertical
direction. In other words, the Ge crystal lattice is subjected to
tension stress in the current direction of the spin transistor, and
accordingly, the electron mobility is effectively made higher. In
this embodiment, the nonmagnetic layer 82 is provided between the
ferromagnetic layer 81 and the semiconductor crystalline layer 84.
If the nonmagnetic layer 82 is a crystalline layer, the nonmagnetic
layer 82 is also affected by the base layer 75 via the
ferromagnetic layer 81. Therefore, the Ge crystal lattice of the
semiconductor crystalline layer 84 becomes slightly larger than the
lattice of the SiGe crystalline layer serving as the base layer
75.
[0069] In a case where the spin transistor is a p-type spin
transistor, SiGe crystals having a lower Ge composition ratio are
preferably used as the material of the semiconductor crystalline
layer 84. With this arrangement, the crystal lattice of the
semiconductor crystalline layer 84 is made smaller than the lattice
of the SiGe crystalline layer serving as the base layer 75, and is
subject to compression stress in the current direction of the spin
transistor.
[0070] As the gate insulating film 88a, a high-dielectric material
having a metal such as Zr, Hf, or La solid-solved in SiO.sub.2 such
as Hf silicate or Zr silicate may be used, instead of a SiO.sub.2
film used in a conventional MOSFET. As the gate electrode 86,
Poly-Si or Poly-SiGe that is p-type or n-type doped is used.
However, in a case where this is combined with a high-dielectric
insulating film, it is possible to use a metal-based material such
as TiN, TaN, TaC, a rare-earth metal, or a rare-earth transition
metal alloy.
[0071] In the third embodiment, the nonmagnetic layer 82 and the
nonmagnetic layer 91 are provided at the interfaces between the
semiconductor crystalline layer 84 to be the channel region and the
source electrode 80 and the drain electrode 90, respectively.
However, as in a modification illustrated in FIG. 14, a source
electrode 80A and a drain electrode 90A from which the nonmagnetic
layer 82 and the nonmagnetic layer 91 are eliminated may be
used.
[0072] As described above, according to the third embodiment and
the modification, writing can be performed with unidirectional
currents, and there is no need to use a transistor for switching
the directions of currents, as in the second embodiment.
Accordingly, the circuit area can be prevented from increasing.
[0073] In the third embodiment, a verify write method may be used
to determine whether desired writing has been correctly performed,
as in the first embodiment. If the desired writing has not been
performed, writing may be performed with a greater write pulse
width, or writing may be performed several times, so as to increase
the probability that the desired writing is properly performed.
Alternatively, depending on the result of reading, writing may be
again performed with a higher or lower write current. For example,
if writing cannot be performed with the current I3, writing is
again performed with a higher current than the current I3. If
writing cannot be performed with the current I1, writing is again
performed with a lower current than the current I1.
[0074] In the above described first through third embodiments and
their modifications, the following materials may be used.
[0075] First, as for the semiconductor layer, an n- or p-type
silicon substrate can be used. Further, it is possible to use a Ge
substrate, a Si.sub.xGe.sub.1-x (0<x<1) substrate, a group
III-V or group II-VI compound semiconductor layer, a magnetic
semiconductor layer, or the like. It is also possible to use a
semiconductor formed on an insulator (such as SOI, GOI, SiGe On I,
III-V On I, or II-IV On I).
[0076] Each of the ferromagnetic layers should preferably have
unidirectional anisotropy. The film thickness of each of the
ferromagnetic layers is preferably in the range of 0.1 to 100 nm.
Furthermore, each of the ferromagnetic layers needs to have such a
thickness as not to be superparamagnetic, and preferably has a
thickness of 0.4 nm or greater. As for the material of each of the
ferromagnetic layers, it is possible to use a Heusler alloy such as
Co.sub.2FeAl.sub.1-xSi.sub.x (0<x.ltoreq.1) or
Co.sub.2Mn.sub.1-xFe.sub.xSi (0<x.ltoreq.1).
[0077] As for each of the ferromagnetic layer, it is also possible
to use Co, Fe, Ni, or an alloy of those materials, an alloy
containing at least one element selected from the group consisting
of Co, Fe, Ni, Ru, Rh, Pd, La, Ce, Pr, Nd, Pm, Sm, Eu, Gd, Tb, Dy,
Ho, Er, Tm, Yb, and Lu, an alloy such as Co--Pt, Co--Fe--Pt,
Fe--Pt, Co--Fe--Cr--Pt, Co--Cr--Pt, Co--Pd, NiMnSb, Co.sub.2MnGe,
Co.sub.2MnAl, Co.sub.2MnSi, or CoCrFeAl, or a magnetic material
formed from the group of magnetic semiconductors such as GeMn,
SiCNi, SiCMn, SiCFe, ZnMnTe, ZnCrTe, BeMnTe, ZnVO, ZnMnO, ZnCoO,
GaMnAs, InMnAs, InMnAb, GaMnP, GaMnN, GaCrN, AlCrN, BiFeTe, SbVTe,
PbSnMnTe, GeMnTe, CdMnGeP, ZnSiNMn, ZnGeSiNMn, BeTiFeO, CdMnTe,
ZnMnS, TiCoO, SiMn, and SiGeMn. Further, a nonmagnetic element such
as Ag (silver), Cu (copper), Au (gold), Al (aluminum), Ru
(ruthenium), Os (osmium), Re (rhenium), Ta (tantalum), B (boron), C
(carbon), O (oxygen), N (nitrogen), Pd (palladium), Pt (platinum),
Zr (zirconium), Ir (iridium), W (tungsten), Mo (molybdenum), or Nb
(niobium) is added to the above mentioned magnetic material, to
adjust the magnetic properties and adjust various kinds of
properties such as crystalline properties, mechanical properties,
and chemical properties.
[0078] As for the nonmagnetic layers 23, 51, 61, 82, 91, and 93, it
is preferable to use an oxide or nitride of Si, Ge, Al, Ga, Mg, Ti,
or the like, or SrTiO, NdGaO, or the like.
[0079] As for the nonmagnetic layers 27, 65, and 95, it is
preferable to use a metal layer containing at least one element
selected from the group consisting of Cu, Ru, Cr, Au, Ag, Hf, Zr,
Rh, Pt, Ir, Al, La, and Mg, or an alloy or oxide layer of those
materials. It is also possible to use an oxide or nitride of Si,
Ge, Al, Ga, Mg, Ti, or the like, SrTiO, NdGaO, Si.sub.xGe.sub.1-x
(0<x<1), a group III-V or II-VI compound semiconductor, or a
magnetic semiconductor. Further, it is possible to use an oxide or
nitride of Si, Ge, Al, Ga, Mg, Ti, or the like, or SrTiO, NdGaO, or
the like.
EXAMPLES
[0080] The following is a detailed description of examples of the
present invention, with reference to the accompanying drawings.
Example 1
[0081] As Example 1 of the present invention, a method of
manufacturing a spin memory having the structure illustrated in
FIG. 1 is described. The procedures for manufacturing this spin
memory are as follows.
[0082] First, a base electrode (not shown) formed with a stacked
film of Ta (50 nm) and Ru (10 nm) is formed on a silicon substrate
having diodes. The numerical values in parentheses indicate film
thicknesses (the same applies to the later descriptions). A
magnetic layer 21 made of TbFeCo (3 nm), a nonmagnetic layer 23
made of MgO (0.8 nm), a ferromagnetic layer 25 made of GdFeCo (3
nm), a nonmagnetic layer 27 made of Cu (0.9 nm), a ferromagnetic
layer 29 made of GdFeCo (6 nm), and a contact layer (not shown)
made of Ta (150 nm) are sequentially formed on the base electrode
by a sputtering technique, to form a ferromagnetic stacked film
20.
[0083] A mask formed with a resist is formed on the ferromagnetic
stacked film 20 with the use of an ArF stepper device. With the use
of this mask, patterning is performed on the ferromagnetic stacked
film 20 by Ar ion milling, to divide the ferromagnetic stacked film
20 into memory cells. After the mask is removed, a 35-nm thick
protection film made of silicon oxide is formed to protect the side
faces of the ferromagnetic stacked film. A mask that is made of
Ta/Ru and is to be used for patterning the base electrode is
formed. With the use of this mask, etching is performed on the base
electrode by Ar ion milling, to divide the base electrode into the
respective memory cells. After an interlayer insulating film (not
shown) made of silicon oxide is formed, etchback is performed to
expose the contact layer. Cleaning is then performed on the upper
face of the contact layer, and a metal stacked film formed with Ti
(15 nm), Al (300 nm), and Ti (15 nm) is formed by a sputtering
technique. Etching is then performed on the metal stacked film by
RIE (Reactive Ion Etching), to form an upper electrode.
[0084] A spin-injection magnetization reversal was tried on the
spin memory manufactured through the above-described process. The
specific procedures are as follows.
[0085] First, a write current J1 was caused to flow at a current
density of 8.times.10.sup.5 A/cm.sup.2 with respect to the junction
area. A read current J2 was then caused to flow at a current
density of 5.times.10.sup.4 A/cm.sup.2, and the device resistance
was measured. As a result, the device resistance was 11
.OMEGA..mu.m.sup.2, converted into RA (Resistance-Area product).
After a write current J3 was caused to flow at a current density of
5.times.10.sup.6 A/cm.sup.2 with respect to the junction area, the
read current J2 was then caused to flow at a current density of
5.times.10.sup.4 A/cm.sup.2, and the device resistance was
measured. As a result, the device resistance was 32
.OMEGA..mu.m.sup.2, converted into RA. The currents J1, J2, and J3
flowed in the same direction.
[0086] As is apparent from the above results, writing of a
low-resistance state and a high-resistance state can be performed
on the spin memory of Example 1 with the unidirectional
currents.
Example 2
[0087] As this example, a method of manufacturing a spin transistor
having the structure illustrated in FIG. 8 is described. The
procedures for manufacturing this spin transistor are as
follows.
[0088] First, a device isolating region (not shown), a gate
insulating film 45, and a gate electrode 47 are formed by a
conventional CMOS process. Ion implantation and activation
annealing are then performed to form a source region 43a and a
drain region 43b. After that, sidewalls (not shown) made of an
insulator are formed on the side faces of the gate electrode 47. A
nonmagnetic layer (a tunnel barrier layer) made of MgO (0.6 nm), a
magnetic layer made of TbFeCo (3 nm), a nonmagnetic layer (a tunnel
barrier layer) made of MgO (0.8 nm), a magnetic layer formed with
GdFeCo (3 nm), Cu (0.9 nm), and GdFeCo (6 nm), and a contact layer
(not shown) made of Ta (150 nm) are sequentially formed by a
sputtering technique, to form a ferromagnetic stacked film.
[0089] A mask formed with a resist is formed on the ferromagnetic
stacked film with the use of an ArF stepper device. With the use of
this mask, patterning is performed on the ferromagnetic stacked
film by Ar ion milling, and only ferromagnetic stacked films 60A
and 60 are left on the source region and the drain region. After an
interlayer insulating film (not shown) made of silicon oxide is
formed, etchback is performed to expose the upper face of the
contact layer. Cleaning is then performed on the upper face of the
contact layer, and a metal stacked film formed with Ti (15 nm), Al
(300 nm), and Ti (15 nm) is formed by a sputtering technique.
[0090] Etching is then performed on the metal stacked film by RIE,
to form an upper electrode (not shown) on each of the source
electrode 60A and the drain electrode 60.
[0091] To simplify the manufacture of devices in this example, the
ferromagnetic stacked films 60A and 60 having tunnel junctions are
formed as the source electrode and the drain electrode or a part of
the source electrode and a part of the drain electrode on the
source region and the drain region of the transistor,
respectively.
[0092] A spin-injection magnetization reversal was tried on the
spin transistor manufactured through the above-described process.
The specific procedures are as follows.
[0093] First, a voltage is applied to the gate electrode, to
energize the channel. A write current is then caused to flow
between the ferromagnetic layer 62 of the ferromagnetic stacked
film 60 and the ferromagnetic layer 62A of the ferromagnetic
stacked film 60A. At this point, the amount of current J1 caused to
flow was 8.times.10.sup.5 A/cm.sup.2, converted into the current
density with respect to the junction area of the ferromagnetic
stacked film 60 (hereinafter all current densities will be defined
with respect to the junction area of the ferromagnetic stacked film
60). A read current J2 was then applied at a current density of
5.times.10.sup.4 A/cm.sup.2, and the resistance between the
ferromagnetic layer 62A of the ferromagnetic stacked film 60A and
the ferromagnetic layer 62 of the ferromagnetic stacked film 60 was
measured, which was 101.OMEGA.. After a write current J3 was caused
to flow at a current density of 5.times.10.sup.6 A/cm.sup.2, the
read current J2 was caused to flow at a current density of
5.times.10.sup.4 A/cm.sup.2, and the device resistance was
measured. As a result, the device resistance was 152.OMEGA.. The
currents J1, J2, and J3 flowed in the same direction.
[0094] As is apparent from the above results, writing of a
low-resistance state and a high-resistance state can be performed
on the spin transistor of this example with the unidirectional
currents.
[0095] The present invention is not limited by the above described
embodiments and examples. For instance, the type of the channel is
not necessary an enhanced type, but may be a depletion type. To
form the channel, not only ion implantation but also a
hetero-interface growing process using modulation doping may be
performed.
[0096] Also, an integrated circuit may be formed with spin memories
or spin transistors of any of the above embodiments and examples. A
memory may be formed with memory cells that include spin memories
or spin transistors of any of the above embodiments and examples,
field-effect transistors, and other memory devices. For example, a
DRAM (Dynamic Random Access Memory) can be formed with those spin
memories or spin transistors in combination with dielectric
capacitors, a FRAM (Ferroelectric Random Access Memory) can be
formed in combination with ferroelectric capacitors, and a MRAM
(Magnetic Random Access Memory) can be formed in combination with
magnetoresistive devices. Further, the spin transistors of the
above embodiments and examples can be used as transistors of an
EEPROM (a programmable read-only memory that electrically performs
data writing and erasing).
[0097] Also, it is possible to provide a memory that uses the
variable magnetization electrode of a spin transistor of the above
embodiments and examples as a storage unit, stores data as the
magnetization direction of the electrode, and reads out the data by
using the magnetoresistive effect caused between the fixed
magnetization electrode and the variable magnetization
electrode.
[0098] The present invention is not limited to the above
embodiments and examples, but modifications may be made to the
components in the implementation phase, without departing from the
scope of the invention. Also, the components disclosed in the
embodiments and examples may be appropriately combined to form
various inventions. For example, some components may be eliminated
from the components disclosed in the embodiments and examples.
Further, some components from one embodiment or example may be
combined with some components from another embodiment or example,
as needed.
[0099] While certain embodiments have been described, these
embodiments have been presented by way of example only, and are not
intended to limit the scope of the inventions. Indeed, the novel
methods and systems described herein may be embodied in a variety
of other forms; furthermore, various omissions, substitutions and
changes in the form of the methods and systems described herein may
be made without departing from the spirit of the inventions. The
accompanying claims and their equivalents are intended to cover
such forms or modifications as would fall within the scope and
spirit of the inventions.
* * * * *